Physics-informed model improves close collision detection for robot arms

A Physics-Informed Collision Learning Framework for Collaborative Robot Motion Generation

Robotics

Summary

Robots with multiple arms need to avoid bumping into themselves when moving closely together. The authors created a new way to predict how close different robot parts are to each other using both physics rules and learned information. This method produces smooth and accurate distance estimates that can be used during robot motion planning to avoid collisions in real time. They tested it on a real robot setup and showed it works well for fast and precise arm movements.

What this means in practice

  • For robotics engineers: Generate collision-aware motion plans for multi-arm robots that operate in close proximity with smoother and more accurate clearance estimates.
  • For industrial automation integrators: Enable safer robot collaborations by integrating real-time collision distance models into dual-arm manipulation tasks, improving throughput and reducing trips.

Authors

Chen Cai, Steven Liu

Abstract

Close-proximity multi-arm manipulation requires collision models that are both geometrically accurate and differentiable enough for real-time optimization. Classical geometry checkers provide reliable distances but are difficult to use inside gradient-based model predictive control, while conservative proxy models can restrict tightly coupled motion. We present PI-UDF, a physics-informed unified differentiable framework for body-to-body collision distance prediction between articulated robots. PI-UDF combines analytical forward kinematics with learnable link-geometry embeddings and a shared residual network to predict pairwise inter-arm distances directly from robot configurations. To improve safety-critical fidelity, we combine quota-driven boundary mining with an asymmetric boundary-crossing penalty that emphasizes false-safe sign errors near the collision boundary. The learned distance field is integrated into nonlinear MPC as a differentiable inter-arm clearance term. We validate the framework on a real dual-Franka platform through high-speed close-proximity 14-DoF dual-arm swapping, sustained single-arm dynamic evasion, and dynamic-evasion planning configurations with frozen, predicted, and target-switching treatments of the moving arm. Hardware experiments and offline Drake/FCL replay show that PI-UDF provides a differentiable inter-arm clearance estimate suitable for closed-loop collision-aware collaborative robot motion generation.